To biopsy or not to biopsy: More than an ultrasound is needed for postmenopausal bleeding evaluation

In April 2026, the American College of Obstetricians & Gynecologists (ACOG) provided a revised Clinical Practice Update recommending that most patients who present with post-menopausal bleeding have an initial evaluation with both transvaginal ultrasonography and endometrial tissue sampling (Obstet Gynecol, 2026). In this update, transvaginal ultrasonography alone is only recommended for patients with one episode of post-menopausal bleeding, a fully visualized endometrium ≤4 mm, with no risk factors for endometrial cancer (which includes exogeneous estrogen use, selective estrogen receptor modulator use, body mass index >30 kg/m², nulliparity, genetic predisposition, Black race, and diabetes mellitus), and without barriers for prompt re-evaluation if needed. This update provides appropriate course correction for multiple identified issues with the prior guidance, including the rising incidence and mortality of endometrial cancer (SEER*Explorer, 2026), the limitation that prior studies on endometrial thickness and endometrial cancer risk were performed largely in non-American cohorts (Gull et al., 2003; Wong et al., 2016), inconsistent data on endometrial thickness in high-risk endometrial cancer subtypes, and disproportionate incidence and delayed diagnosis in non-Hispanic Black populations (Wright and Havrilesky, 2026).

Despite these updated recommendations, the criteria for transvaginal ultrasonography alone still leave room for missed cancer diagnoses as identified by this present study. Liu et al. utilized data from their PostMenopausal Bleeding Assessment (PUMBA) study to evaluate the diagnostic performance of transvaginal ultrasound across different racial and ethnic backgrounds and to identify factors associated with suboptimal ultrasound-based diagnostic follow-up. They compared the endometrial visibility and thickness of TVUS of endometrial cancers of endometrioid versus non-endometrioid histology with attention to patient race, ethnicity, and the presence of fibroids. Only patients who had a tissue diagnosis of endometrial cancer within 3 months after the initial transvaginal ultrasound were included. Thus, this study includes ultrasound findings from patients who had adequately visualized endometrium and had the potential to be missed by the new ACOG guidelines (Liu et al., 2026). This retrospective cohort study included 185 patients from 3 clinical sites (recruited as part of the larger multisite PUMBA study) diagnosed with endometrial cancer within 3 months of a transvaginal ultrasound for postmenopausal bleeding. Among this cohort, which included 7.0% Hispanic, 27.0% non-Hispanic Black (NHB), and 56.8% non-Hispanic White (NHW) patients, incomplete endometrial visualization occurred in 18.9% of the overall cohort and was numerically more common in Hispanic and NHB patients, in non-endometrioid cancers, and in patients with fibroids. When considering all patients in the cohort, 20.5% had either incomplete endometrial visualization or a falsely thin endometrial stripe. When analyzing only patients with complete endometrium visualization, the mean endometrial thickness was well over 4 mm, but 5.6% of cancers had a ≤4 mm endometrial stripe. This did not differ based on histology (endometrioid 5.5% vs non-endometrioid 6.1%) or race/ethnicity.

Multiple findings from this study support the need for more evaluation than an ultrasound. As the authors point out, there is a higher false negative rate for endometrial thickness cutoff based on transvaginal ultrasonography alone than previously cited (Long et al., 2020). Non-visualization of the endometrium was common, especially in non-endometrioid histologies, a limitation which is overcome by recommending upfront endometrial sampling in combination with ultrasound. Non-Hispanic Black patients more commonly had inadequate endometrial visualization and fibroids compared to NHW patients, and these patients were also more likely to be diagnosed with non-endometrioid cancers, which reinforces the limits of ultrasound alone in this population and the consideration of Black race as a risk factor for endometrial cancer development (Nolin et al., 2026; Doll et al., 2021).

This study raises several interesting additional considerations. As the authors astutely discuss, over 5% of endometrial cancers in their study had a thin endometrial stripe, which would miss a significant proportion of endometrial cancers on a population basis. While this >5% rate includes NHB patients who would be considered high risk in the current guidelines, based on provided supplementary data, at least 4.2% of non-NHB patients had adequate endometrial visualization with a ≤4 mm endometrial stripe. This is still an overall high rate of missed diagnoses based on ultrasonography alone. Hispanic and patients classified as “Other” or with missing race more commonly had an incompletely visualized endometrium and fibroids compared to NHW patients. While we cannot ascertain how many patients would have been considered at high risk of endometrial cancer based on medication use, BMI, nulliparity, genetic predisposition, or diabetes and while the numbers are small, this warrants close future attention to ensure that other potentially high-risk patients are not being left behind in screening guidelines, or that guidelines are updated to include a broader definition of high-risk patients as updated information becomes available.

The present study included patients diagnosed with cancer based on endometrial biopsy, dilation & curettage, or hysterectomy. It is not described why patients with an initially thin endometrial stripe ultimately had sampling or surgery. Many of these patients may have had persistent or recurrent post-menopausal bleeding which would have necessitated additional testing, but it is also possible that the false negative rate quoted by the authors is an underestimate of endometrial cancer diagnoses in an all-comer population undergoing ultrasound for post-menopausal bleeding. This study population, by definition, had follow-up after their initial transvaginal ultrasonography, which could also lead to an underestimation of the false negative ultrasound results in a general population, who would be expected to have a proportion of patients lost to follow-up.

It is also possible that some of the study patients were unable to tolerate an initial attempt at an endometrial biopsy and thus required subsequent additional procedures or surgery to obtain a diagnosis during the 3-month follow-up period. It will be critical to continue responding to patient requests for improved analgesia for office gynecologic procedures as more patients are recommended for endometrial sampling based on the updated guideline. Additionally, since up to 10% of in-office endometrial sampling fails (van Hanegem et al., 2016), it will be important in future studies to examine the burden of cost and time required for additional procedures for patients presenting with post-menopausal bleeding who cannot have an office endometrial biopsy. It will be equally important to ensure that barriers to necessary additional procedures are mitigated to avoid the delays in diagnosis addressed by the updated guidelines.

While all patients in this study were >50 years old, this study does not by inclusion criteria strictly examine post-menopausal bleeding for which the new ACOG guidelines apply. As endometrial cancer and endometrial intraepithelial neoplasia continue to rise in young patients (SEER*Explorer, 2026; Guo et al., 2021; Liu et al., 2023), non-menopausal populations will warrant additional scrutiny in terms of when and how often endometrial sampling is recommended to avoid missed cancer diagnoses, especially in patients with adequately visualized, thin endometrial stripes.

Overall, this study provides important reinforcement to perform both transvaginal ultrasonography and endometrial sampling in most patients with post-menopausal bleeding. This is particularly important given adequate endometrial visualization could still be associated with 5% missed diagnosis rate, as in the study patients with a ≤4 mm thin endometrium that correlated with a biopsy-positive cancer. It also provides additional nuances in terms of how ultrasound visualization, fibroid status, race, ethnicity, and ultimate cancer diagnosis are associated. These data overall support expanded indications for upfront endometrial sampling and are useful for patient counseling. As individual providers and healthcare systems alike adjust to the recommendation for expanded upfront endometrial sampling in conjunction with ultrasound for postmenopausal bleeding, it will be important to ensure equitable implementation of guidelines and ongoing data collection to evaluate how cancer outcomes are affected.